Identifying malnutrition in emergency general surgery: systematic review

Abstract Background Emergency general surgery practice is high risk. Surgery is a key part of treatment, with resultant catabolic stress and frequent need for nutritional support. The aim of this study was to examine the current methods of defining and determining malnutrition in emergency general surgery. This included examining the use of nutrition screening and assessment tools and other measures of malnutrition. Methods MEDLINE, Embase, Cumulative Index to Nursing and Allied Health Literature, trial registries, and relevant journals published between January 2000 and January 2022 were searched for studies of adult patients with any emergency general surgery diagnosis, managed conservatively or operatively, with an assessment of nutritional status. Mixed populations were included if more than 50 per cent of patients were emergency general surgery patients or emergency general surgery results could be separately extracted. Studies in which patients had received nutritional support were excluded. The protocol was registered with PROSPERO, the international prospective register of systematic reviews (CRD42021285897). Results From 6700 studies screened, 324 full texts were retrieved and 31 were included in the analysis. A definition of malnutrition was provided in 23 studies (75 per cent), with nutritional status being determined by a variety of methods. A total of seven nutrition screening tools and a total of nine ‘assessment’ tools were reported. To define malnutrition, the most commonly used primary or secondary marker of nutritional status was BMI, followed by albumin level. Conclusion Wide variation exists in approaches to identify malnutrition risk in emergency general surgery patients, using a range of tools and nutrition markers. Future studies should seek to standardize nutrition screening and assessment in the emergency general surgery setting as two discrete processes. This will permit better understanding of malnutrition risk in surgical patients.


Introduction
Emergency general surgery (EGS) is the single largest group of surgical patients admitted to hospital in the UK 1 .EGS represents 4.2 per cent of all hospital admissions in the UK and 7.1 per cent in the USA 2 .More than a quarter of patients admitted in this setting require surgery 2,3 , which carries a high risk of morbidity and mortality [4][5][6] .Malnutrition is an established risk factor for worse outcomes, including more frequent nosocomial infections 7 , extended length of stay 6 , increased mortality 5 , and increased healthcare service costs 6,8,9 .
Despite the size of the challenge, there is no universally accepted definition for malnutrition in the literature or in EGS clinical guidelines [10][11][12][13][14][15][16] .The current incidence of malnutrition in EGS patients has not been clearly established, but may be as high as 40-60 per cent 5,17 .Knowing which patients are malnourished will allow targeted nutritional support, which may improve outcomes.
The identification of malnourished patients is typically a two-step process 18,19 .This involves an initial screening, followed by a more comprehensive assessment by an appropriate healthcare professional for those deemed to be at risk.There is an abundance of methods and tools to determine the risk of malnutrition in patients, though very few are validated in the EGS cohort.There is no international consensus on which screening tool should be used 12,15,20 .
The Malnutrition Universal Screening Tool (MUST) is one such screening tool that has been validated across multiple healthcare settings and populations.It was developed from a community tool by the Malnutrition Advisory Group of the British Association for Parenteral and Enteral Nutrition (BAPEN) 21,22 and consists of three domains (body mass index (BMI), unintentional weight loss, and acute disease or lack of nutrition affect).A score of two or more out of six indicates a high risk of malnutrition and the need for specialist dietetic involvement.It is widely implemented across UK settings 23,24 and is used in the majority of UK health institutions as the standard screening tool.
The aim of this study was to perform a systematic review of observational studies to describe how malnutrition is identified in EGS.It focused on nutrition screening and assessment tools, and their components, as well as other markers of malnutrition.

Protocol and registration
The systematic review was performed with reference to the Cochrane Handbook and reported using the PRISMA https://doi.org/10.1093/bjsopen/zrad086Systematic Review guidelines 25 .The protocol was registered with PROSPERO, the international prospective register of systematic reviews (CRD42021285897).

Eligibility criteria
Studies of adult (aged 18 years and older) patients with any EGS diagnosis, managed conservatively or operatively, were included (Table 1).Participants must have had their nutritional status determined at some point in their hospital stay using a screening tool, assessment tool, or other measure of malnutrition.Studies on mixed patient populations (mixed surgical specialty, emergency and elective general surgical, or surgical and medical) were included if the EGS and non-EGS results were reported separately or if 50 per cent or more of the population reported were EGS patients.The same applied to studies of EGS and trauma patients.Both randomized and non-randomized studies were included, as the extracted data related to a descriptor, and were presented in a narrative manner.However, studies where a nutritional intervention implemented at any stage (oral, enteral, or parenteral) was described in the population were excluded, as observational studies were considered more likely to reflect real-world practice.Additional exclusions applied to studies of pregnant women, studies of patients with eating disorders, inappropriate study design (letters to the editor), review articles, and abstracts with no corresponding full text.Whilst studies published from any country were considered, non-English texts were excluded, as no resource was available to secure accurate and reliable translations.

Nutrition screening tool
This was defined as a tool to identify individuals at risk of malnutrition.Although this would ordinarily subsequently lead to a full nutritional assessment by a dietitian, this was not necessary to be included in the review.

Nutrition assessment tool
Due to a lack of clear terminology in the literature, in this review a nutrition assessment tool was defined as a method that: characterized the nutritional status of an individual to confirm whether they were malnourished and/or to what degree; had 'assessment' in the name of the tool and was referred to by that throughout the study; or was used when the authors considered that a nutritional assessment (rather than screening) had taken place.

Outcomes
The primary outcome of this review was to identify which nutrition screening or assessment tools are used in the EGS population and what their component parts are.
Secondary outcomes included mapping the range of alternative criteria used in defining malnutrition in EGS.This was purposefully kept broad to allow a full overview of the range of methods currently used to assess nutritional status.It included anthropometric indices (for example height, weight, BMI, and skin-fold thickness), biochemical markers (for example albumin and lymphocyte count), and others (for example wound healing and functional recovery).

Search strategy
A systematic search for studies from January 2000 to January 2022 (sample search strategy available in the Supplementary material) was performed within the following databases: MEDLINE (via Ovid SP), Embase (via Ovid SP), Cumulative Index to Nursing and Allied Health Literature (CINAHL) (via EBSCO), and Cochrane CENTRAL.In addition, the following clinical trial registries were also searched: ClinicalTrials.gov, the WHO International Clinical Trials Registry Platform, and the EU Clinical Trials Register.Reference lists of eligible studies were checked, as were reports published by the relevant organizational bodies, including the Department of Health UK, the National Institute for Health and Care Excellence (NICE), the Royal College of Surgeons, and the British Society of Parenteral and Enteral Nutrition.Selected journals were searched, including Clinical Nutrition, the Journal of Parenteral and Enteral Nutrition, the Journal of Human Nutrition and Dietetics, Nutrition and Dietetics, and BMC Nutrition.

Study selection and screening
Titles and abstracts were screened for eligibility by two independent reviewers (D.L.A. and A.R.) using Rayyan 26 .Manuscripts meeting eligibility, or where there was insufficient detail to make an assessment of eligibility, were obtained for full-text review.Disagreements, where necessary, were settled with a third reviewer (M.J.L.).All reasons for exclusion of studies were documented.

Data extraction
A standardized data extraction sheet was used to extract details of relevant studies using Microsoft Excel (version 2010).

Data items
The data to be extracted aimed to meet the outcomes as outlined above.This was performed by D.L.A. and validated by M.J.L. Disagreements were settled by majority decision between D.L.A, A.R., and M.J.L. Missing data were recorded as such.Duplicates were removed.Data items of interest included patient and disease characteristics, setting, and nutrition tools used.Markers of malnutrition were also identified.

Data synthesis
A narrative synthesis was performed.This included reporting the frequency of nutrition tools used, the settings in which they were used, and the patient population included in the study.Included tools were reviewed to identify component items.These were mapped to ascertain 'core areas' of malnutrition used in the literature.Given this review reports a summary of how nutritional status is assessed rather than the clinical outcomes of patients, no bias assessment was performed for the studies included.This was not deemed necessary, in keeping with other similar reviews 27 .

Search results
The searches identified 7243 records, from which 543 duplicates were removed.Of the remaining 6700 studies screened for eligibility, 6374 studies were excluded based on title and abstract.A further two could not be retrieved.Full-text review of 324 studies was performed.The majority of these were conference abstracts with no corresponding full study (Fig. 1).There were 31 studies eligible for inclusion.

Definitions for malnutrition
Of the 31 studies included in this review, 23 studies (75 per cent) provided a definition of malnutrition.This was by use of a nutrition screening tool alone 33,34,39,48,50 , an assessment tool 28,29,32,37,38,[40][41][42]51 , or a nutrition marker, either on its own 35,43,45 or combined 5,31,44,49,52,53 . Thre were two studies that provided no definition of malnutrition 46,54 .In a further six studies the definition was not explicit due to an absence of laboratory value cut-offs 55 , an absence of scoring system cut-offs 56 , an absence of nutrition marker cut-offs 36,47,57 , or the tool used was for another reason (frailty) and malnutrition was not clearly defined within that tool or elsewhere in the study 58 .A total of three studies used a nutritional screening tool, followed by an assessment 5,30,48 , although the screening tool was not stated in one study 5 .Overall, BMI was the most commonly used criterion to define malnutrition as either a primary or secondary marker of nutritional status, or as a key component of a nutritional tool, in 20 (64.5 per cent) studies. Serum alumin was used as the same in 18 (58.1 per cent) studies.
The range of criteria for malnutrition according to a 'core area' is mapped in Fig. 2.These core areas are history, clinical examination, blood (laboratory) values, disease criteria, formulae, and functional tests.Core areas have underlying themes, which overlap with one another, representing the multiple criteria some studies used to define malnutrition (Table S7).

Nutrition screening tools
Nutritional status was determined using a variety of methods, but each tool was used by one paper only (Table 3).There were seven nutrition screening tools or methods in use: the Hong Kong Chinese Malnutrition Screening Tool, the Malnutrition Screening Tool, hand-grip strength, the 3-Minute Nutrition Screening Tool, the Geriatric Nutritional Risk Index, the Trauma and Emergency General Surgery Frailty Index, and the Canadian Nutrition Screening Tool.

Nutrition assessment tools
There were nine nutrition assessment tools, of which the most commonly used was the Subjective Global Assessment (SGA) tool.A summary of these tools is presented according to year published (Table 4).Nutrition screening and assessment tools according to 'core areas' are also presented (Table 5).

Alternative nutrition markers (combined measures)
There were seven combined measures used to define malnutrition between nine studies 5,31,34,36,44,47,49,52,53 .Hypoalbuminaemia was a key factor in eight of these.The combined measures included: appetite and weight loss; albumin and weight loss; albumin, BMI, or weight loss; albumin, BMI, and weight loss; albumin or BMI; a complex combined measure of weight loss, underweight based on ideal weight, muscle wasting, and inadequate protein-energy intake; and, BMI, arm circumference, and skin-fold thickness (Table 7).

Implementation of a nutritional tool
Patients' nutritional status was assessed at a number of time points.Almost half of the studies (15 studies (48.4 per cent)) performed the assessment at admission or before surgery (Table 2).

Discussion
This systematic review reports the current methods used to identify malnutrition in EGS.Within the 31 included studies, a range of approaches were used.These included seven nutrition screening tools, nine nutrition assessment tools, seven single nutrition markers, and seven combined measures used to define malnutrition.There is considerable heterogeneity in 'core areas' in the tools used in EGS research and their components.
The findings in this review likely reflect the inconsistency among guidelines and the absence of a universal definition for malnutrition.NICE suggests MUST to screen for malnutrition in the UK and this is widely adopted 12 .The European Society for Clinical Nutrition and Metabolism (ESPEN) recognizes a number of tools that can be used in a hospital setting.These include the Nutritional Risk Score 2002 (NRS 2002), the Mini Nutritional    37 , 2017 Formula used in study: NRI = 1.519 × serum albumin (g/l) + (0.417 × actual weight/ideal weight) Original Buzby Index: NRI = (15.9×serumalbumin, g/l) + (0.417× % usual body weight) Where % usual body weight is current weight as a % of usual weight either 2 or 6 months ago 59 Mild malnutrition = >97.5 Moderate malnutrition = 83.5-97.5  Severe malnutrition = <83.5

Mini Nutritional Assessment Short Form (MNA-SF)
Fuertes-Guiró et al. 51    Assessment Short Form (MNA-SF) and MUST 15 .ESPEN's guidelines on nutrition in critically ill patients advise that a general clinical assessment of nutrition should be performed until a specific tool has been validated, although every patient in an ICU for more than 48 h should be considered high risk for malnutrition 60 .Conversely, the Academy of Nutrition and Dietetics and the American Society for Parenteral and Enteral Nutrition (ASPEN) advise using nutritional risk screening tools such as NRS 2002 or Nutrition Risk in Critically Ill (NUTRIC) 20 .The Global Leadership Initiative on Malnutrition (GLIM) guidelines state that screening should be performed using a validated screening tool and, although none is recommended, tools such as NRS 2002, MNA-SF, MUST, and SGA are presented 14 .The identification of malnutrition is a two-step process; screening of patients who may be at risk of malnutrition is followed by an assessment to confirm the diagnosis, categorize, and plan treatment.This is recognized by the British Dietetic Association 18 and overseas by the Academy of Nutrition and Dietetics 19 , as well as a number of guidelines.The nutritional assessment is typically performed by a trained professional 18,19 .GLIM endorse that phenotypic (weight loss; BMI; muscle mass) and aetiological (food intake or assimilation; inflammation or disease burden) criteria can be used for a diagnosis of malnutrition, although consultation with a trained professional is recommended for a comprehensive assessment 14 .Only three studies used a nutritional screening tool, followed by an assessment 5,30,48 , although the screening tool was not stated in one study 5 .Of these three studies, the assessment of malnutrition was done by a dietitian in one study alone 5 .The absence of this two-step process in EGS research likely has two effects.
First, it leads to confusion within the literature, and possibly clinical practice, regarding what is meant by an 'assessment' of nutritional status.Some tools exist as nutrition 'assessment' tools in that a more global assessment is performed.Others are simply named assessment tools, but may be no more complex than a screening tool.Consequently, assessment tools are conflated with screening tools and vice versa.For example, SGA has been evaluated in many studies including of all types of adult inpatients 61 .It has subsequently become considered a gold standard against which other tools are assessed 61 .This is despite it being a subjective process with minimal 'assessment'.
Second, the absence of a two-step process adds to the heterogeneity of EGS research, which is already a labyrinth of complex diagnostic, investigative, and management processes.This may further hamper the development of comparable studies and translation of malnutrition identification into routine practice.Of the nine 'assessment' tools in this review, three are based on formulae (Prognostic Nutrition Index (PNI), Nutrition Risk Index (NRI), and modified Glasgow Prognostic Score (mGPS)), two are subjective (SGA and Patient-Generated SGA (PG-SGA)) and three are aimed at identifying malnutrition within the framework of frailty (Mini Nutritional Assessment (MNA), MNA-SF, and Geriatric 8 (G8)).
Finally, albumin has often historically been regarded as a marker of nutritional status 62 and there remains a degree of controversy with respect to its role in assessing malnutrition 63,64 .This study found that albumin was used in more than half of studies as a primary or secondary marker of malnutrition.It clearly has a role in the inflammatory state as a negative acute phase marker, the postoperative stress response, and determining disease severity; all of which may impact nutrition.Moreover, hypoalbuminaemia is associated with increased mortality and morbidity 65,66 and waiting for this to normalize before surgery reduces these risks 67 .However, there is inadequate evidence for its direct role in determining nutritional status 68 .Indeed, it does not help in deciding which patients should receive nutritional support 69 and a normal albumin level can be seen despite severe malnutrition 70 .
A number of international bodies advise against the use of albumin as a marker for malnutrition.NICE recommends that Fernandes et al. 43 , 2019 Malnutrition: albumin ≤3.5 g/dl Krishna et al. 45 , 2019 Preoperative albumin (reference value 3.5-5.5 g/dl) was used as an indicator of the nutritional status of the patients Significant hypoalbuminaemia (that is malnutrition) = 3.0 g/dl Takano et al. 40 , 2021 Preoperative levels (cut-off NS, although Table 3 28 , 2017 TNL <1200 in 1 mm 3 of blood To determine the weakening of resistance that accompanies malnutrition, the TNL was determined by: TNL = (% of lymphocytes × number of lymphocytes)/100 NS, not specified; TNL, total number of lymphocytes.
low albumin is more likely to be a reflection of acute disease rather than malnutrition 12 .ESPEN support this, but reiterate that hypoalbuminaemia is a surgical risk factor in its own right 60,71 .ASPEN categorically states that albumin and pre-albumin should not be used as nutrition markers 72 and GLIM states that albumin is a proxy marker of inflammation only 14 .Although albumin may have a role in prognostication for surgical outcomes, the use of albumin as a marker of the current state of malnutrition should be abandoned, clinically and in future studies.
It is important to recognize, however, that a patient's nutritional status is a dynamic process and, as such, so should be the process of identifying it.To begin with, there is the patient's pre-existing nutritional status.They may be underweight, exhibit a degree of chronic malnutrition, or, ever-increasingly in the Western world, be overweight or obese.The subsequent treatment and management of their diagnosis, as well as future possibilities such as a persisting ileus or reduced oral intake, will also affect a patient's nutritional status.Whilst almost half of the studies determined nutritional status at admission or before surgery, it was not stated in close to 40 per cent of studies and certainly many studies did not comment on whether patients were reassessed (data not stated).
This systematic benefits from exploring a research question with direct relevance to a large proportion of acute surgical admissions.It was registered prospectively and had wide inclusion criteria, allowing for studies of adult patients with a general surgical diagnosis to be included, whether managed conservatively or operatively.
This study is not without limitations.Firstly, although more than half of the included studies (18 studies (58.1 per cent)) were confined to EGS patients, mixed patient populations (mixed surgical specialty, emergency and elective general surgical, surgical and medical, or trauma patients) were also eligible for inclusion, with the EGS group extracted.Although consideration was given to limit studies only to those of patients with 'high-risk' EGS diagnoses such as those defined by Symons et al. 73 or those using National Emergency Laparotomy Audit (NELA) database criteria or similar 74 , this would have led to a paucity of data.A total of three studies had fewer than 50 per cent EGS patients 33,41,46 .The inclusion of these studies did not lead to irrelevant nutrition screening or assessment tools that were more targeted to populations other than EGS patients.Ho et al. 33 was one of only three multicentre studies examining adult surgical patients undergoing elective and emergency operations on the luminal gastrointestinal tract.They used C-MUST, which is based on MUST and has been extensively validated and is widely used in the UK 21 .The second study, by Waqar et al. 46 , examined patients who had undergone a laparotomy in the elective, emergency, and trauma setting.In this study by Waqar et al. 46 75 .Finally, this review did not correlate how malnutrition was identified with the range of diagnoses, clinical outcomes, or patient-reported outcomes in EGS.This was not in the initial remit of the review, although it was noted that there were inconsistencies in reporting clinical outcomes.In addition to the various study designs, this would have been difficult to interpret and generate meaningful conclusions.
Given that there is a wide variation in study design, population, definitions, and the nutrition tools and markers used to determine malnutrition in EGS, caution must be used when interpreting the findings of these studies and applying them to the bedside.It is imperative to identify malnutrition in EGS, both at admission and throughout the duration of a patient's hospital stay, but the current strategies are inadequate.Minimizing heterogeneity in study design and patient recruitment will be difficult in this setting given its very nature.
The literature identifies many nutrition screening tools and developing another new tool may not be helpful.It is crucial to emphasize screening and assessment as two distinct pathways.Gaining a better understanding as to how clinicians diagnose and manage malnutrition in EGS practice, and why they use the methods they do, may offer useful insights as to how this process can be made more robust and reproducible.It is recommended that future research in EGS populations should report malnutrition risk at baseline using a recognized tool to fully appreciate its impact on outcomes.Standardizing this would help with future data synthesis and interpretation.
Reducing variation will permit better understanding of malnutrition risk in surgical patients.Identifying patients who are malnourished, or at risk of being or becoming so, during their admission will allow targeted nutrition intervention and this may improve surgical outcomes.

Funding
The University of Sheffield Institutional Open Access Fund funded the article processing charges.There were no other sources of funding.

Fig. 2
Fig. 2 Mapping of criteria used to define malnutrition

Table 2 Summary characteristics of the included studies
29lues are n (%).A total of 31 studies were included.*Onestudyisafollow-up study28.†Onestudystated day 4-6, but not clear if this was before or after surgery29.EGS, emergency general surgery.

Table 7 Alternative nutrition markers in emergency general surgery (combined measures)
Ashmore et al. | 9 , no definition of malnutrition is offered, although this was the case in a quarter of studies in this review.Lastly, the study by Welch et al. , in which only 16 patients (19.8 per cent) were EGS patients, assessed the feasibility of conducting acute sarcopenia research.This study by Welch et al. 41 used MNA.It is a well-validated tool for older populations and has questions relating to a range of anthropometric measures, including weight loss and BMI, and a dietary assessment